In AR implementations, a user may observe the actions of an AR element in a real space (e.g., environment) only when the user is viewing a rendering of the element and/or via an intermediate device. For example, the user may need an intermediate display device (e.g., tablet) to observe the result of an interaction between the AR element and the real objects in the real space. Such implementations may lead to unnecessary utilization of resources such as a device, processor, memory, and power (e.g., reduced battery life). In addition, such implementations may be constrained to a particular device where the AR element is computed and/or displayed.